The most important number in Blue Owl Capital’s launch of Kirkwood Infrastructure Group is not the 400 miles of fiber network or the 40 data centers it already controls. It is the 40% decline in Blue Owl’s share price since January and the $4.7 billion in redemption requests its private credit funds absorbed in the second quarter.
Blue Owl is building long-duration infrastructure assets at the same moment its investors are asking for their money back. That tension is the real story.
The firm announced Wednesday that funds it manages have created Kirkwood Infrastructure Group to develop, own, and operate fiber networks serving hyperscalers, data center operators, and community-level networks. The venture has already integrated South Reach Networks, a fiber company Blue Owl acquired last year, and is planning 200 miles of new conduit and high-capacity fiber across Louisiana and Mississippi.
On its face, the move is logical. The AI-driven data center buildout is generating demand for connectivity that will persist for years. Goldman Sachs projects $7.6 trillion in big tech capital expenditure by 2031. Fiber networks are the physical backbone of that compute expansion. Owning the pipes is a defensible long-term bet.
But the capital structure behind this bet deserves scrutiny. Blue Owl’s private credit funds have been the industry’s largest recipients of redemption requests for two consecutive quarters. The firm capped withdrawals in Q2 after $4.7 billion in requests, down only modestly from $5.4 billion in Q1. Its stock has not recovered from a March sell-off tied to broader fears about AI-related concentration risk in private credit.
Fiber infrastructure is not a liquid asset. It requires patient capital with a multiyear horizon. Building conduit and laying cable generates no immediate cash flow. The returns come from long-term lease agreements with tenants who need certainty of connectivity. This is the opposite of the liquidity profile that investors demanding redemptions are seeking.
Blue Owl is effectively asking its limited partners to accept that the firm’s capital is better deployed in illiquid, long-duration infrastructure than returned to them. That is a difficult argument to make when redemption queues are growing, not shrinking.
The firm’s prior data center financing deals show the pattern. In August 2025, Blue Owl partnered with Pimco to invest $29 billion in Meta data centers in Louisiana, with Blue Owl contributing $3 billion. In June, it provided nearly $1 billion to recapitalize a data center project in Northern Virginia. These are large, illiquid commitments that lock up capital for years.
Kirkwood represents a further extension of that strategy. But fiber networks are even less liquid than data center shells. They have no alternative use. Their value depends entirely on the tenant’s need for connectivity and the landlord’s ability to maintain the network. If the AI buildout slows or shifts geography, fiber assets are harder to reposition than a warehouse or an apartment building.
Who benefits from this structure? Blue Owl’s management fees. The firm earns on assets under management, and Kirkwood adds to that base. The hyperscalers and data center operators who need fiber benefit from having a well-capitalized counterparty willing to build speculatively. The existing South Reach Networks team gets a larger platform.
Who is exposed? The limited partners in Blue Owl’s funds. They are funding long-duration infrastructure with capital that may not match the asset’s holding period. If redemption pressure continues, the firm may face a choice between selling assets at a discount or further restricting withdrawals. Neither outcome is favorable for LPs seeking liquidity.
The broader market signal is about capital allocation discipline in the AI infrastructure boom. Private credit has flowed aggressively into data centers and related assets because the demand story is compelling. But the liability side of the balance sheet matters. Funds that offer quarterly liquidity cannot safely own assets that take a decade to mature. The tension between what investors want and what the assets require is the defining risk of this cycle.
Blue Owl is not wrong about the opportunity in fiber. It may be wrong about the capital structure that funds it. The next phase of the market will not be defined by who builds the best network. It will be defined by who controls the capital that can wait for the return.